摘要
通过对自组织数据地图SOM(SelfOrganizingMap)理论及技术的研究和应用,力求找到近海水质分析的新途径,并以渤海湾水质数据为例,采用批处理SOM算法为计算核心,利用MATLAB语言编制的SOM工具箱开发了近海水质分析软件包。该软件包通过SOM的训练对数据特征进行分析和显示,并在此基础上进一步分析SOM训练结果,将监测数据分成5种不同的污染类型,还实现了数据跟踪和新的监测数据的自动归类。应用表明,提出的方法以多种图形形式直观、综合、深入的分析近海水质,可为近海水域污染状况的辨识和采取相应措施提供决策支持。
A new approach to coastal water quality analysis is expected in this paper through the study on the SOM (Self Organizing Map). Firstly, the water quality data of Bohai Bay is ready. Then, a set of software for coastal water quality analysis is developed based on the batch version algorism of the SOM and the SOM toolbox in the MATLAB environment. This software can analyze and display the data characters through the training of the SOM. Furthermore, the training results can be analyzed so that the data is divided to five different kinds of pollution. Lastly, it is also realized that the monitored data serial can be tracked, and the new data can be classified automatically. Through application it can be found that this study helps to analyze the coastal water quality in depth with several kinds of graphics, which supply the decision support to recognize the pollution status and take the corresponding measures.
出处
《水科学进展》
EI
CAS
CSCD
北大核心
2005年第4期569-573,共5页
Advances in Water Science
基金
天津市科技发展计划资助项目(033113811)
天津市自然科学基金资助项目(043606511)
天津大学青年教师基金资助项目(985200540)~~
关键词
自组织数据地图
近海
水质分析
污染类型
<Keyword>self qrganizing map
coastal water
water quality analysis
pollution kinds
